Describe · Build · Export
Campaign AudienceBETA
When a client campaign needs a list, build one for that campaign. Describe who they need to reach in plain language and Claude writes the criteria, tells you exactly how many people match, and shows you ten of them before you decide anything — then hands you a workbook with your client's name on it.
The List Your Client Keeps Asking For
Describe the buyer in a sentence — see exactly who matches, and ten of them, before you decide
A small agency cannot justify a seat on a contact platform for the two campaigns a year that need one. This builds the list per campaign instead: you describe who your client is trying to reach, and Claude turns it into real criteria, tells you precisely how many people match, and puts ten of them in front of you — names, companies, emails, what each business does — while you are still deciding.
“Owner-run marketing and creative agencies in the US, 2 to 30 staff. The founder is the person I want.”
No filter menus. No 93-item industry list to scroll.
Checked against the live list — an invalid one is dropped and named.
About three seconds. A number, not a range.
Real people from that list — and they are the first ten of it, not a preview that disappears. You judge the fit from the actual names, then choose how many to take.
What One Row Actually Holds
Enough to write the first line — not just an address to blast
Most lists give you a name, a title and an email, and leave you to guess the rest. Every record here arrives with the company attached — what it does, how big it is, when it started — which is the difference between a mail merge and a message that sounds like you looked.
Illustrative record — the fields are the real ones. On a live pull of 300, every row carried an email and a company domain, and 296 of 300 carried a description like the one above.
Built For The Agency That Is Not An Enterprise
A list per campaign — instead of a seat you cancel in six weeks
Apollo, Clay and ZoomInfo are built for teams that prospect every day. Most small agencies do not — they need a list when a client campaign needs one, three or four times a year. Paying for a seat all year to use it twice is the problem, and it is the only reason this exists.
A subscription, whether you use it or not
A list built for one campaign, when the campaign happens
The Thing You Actually Hand Over
A workbook with your client's name on it
The client, the audience and Claude's description of the criteria run across the top, so anyone who opens it understands what they are looking at. Every column filters. A second sheet explains what each column means and how complete it is — including what the data does not have.
How agencies work it
Describe who your client needs to reach
Plain language, the way you would brief a colleague — “owner-run marketing agencies in the US, two to thirty staff, the founder is the person I want”. Claude turns that into real criteria: industry, seniority, company size, location, even the software a company runs.
See the size and a real sample
Claude breaks the brief into segments and tells you exactly how many people match each one — a precise number, not an estimate. Then it shows you ten of them: names, titles, companies, emails, what each business actually does. You are looking at the real list before you decide anything.
Pull what you want
Choose a quantity per segment and the list lands in seconds. Pull more from the same segment later and you get new people, never the same ones twice.
Hand it to the client
Download a styled workbook with the client name, the audience and Claude's description of the criteria across the top — or an emails-only CSV that drops straight into a sending tool. It reads like agency work product, not a database dump.
What's inside
Claude writes the criteria
The source offers 93 industries, 206 countries and over 1,600 technologies. Nobody scans that. You describe the buyer in a sentence and Claude picks the values — checked against the live list, so a criterion that does not exist is dropped and named, never silently passed along.
Broken into segments
One brief becomes the demographics an agency would message differently — agency owners, ecommerce operators, practice principals. Each gets its own criteria, its own count and its own sample, so you can buy one and skip another.
Three kinds of audience
Email outreach (business emails, built for a send), direct mail (business mailing addresses for a printed campaign), and tech stack (companies running a specific platform). Picking one changes the form, the guidance and the columns you get — the deliverable follows the job.
An exact count, in seconds
Not “about 30,000” — the precise number of people who match, returned in about three seconds. It reframes the decision from a guess into a choice: there are this many, how many do you want?
Ten real people, up front
Actual names, titles, companies, emails and what each business does — pulled from the live list, not a mock. And they are genuinely yours: those ten are the start of your list, not a throwaway demo.
Reaching your number, visibly
Ask for 40,000 and Claude widens the net to get there — adding adjacent industries, then larger company sizes — and shows you each rung it climbed. If reaching your number would mean changing who you are actually reaching, it stops and says so instead of quietly padding the list.
Rank stays true to the brief
Ask for founders and you get founders. The criteria pin seniority precisely — a live pull came back 100% owners, with no managers, no coordinators and no interns drifting in from a looser filter.
Junk removed up front
Duplicate people under two addresses, and addresses the source itself has flagged, are stripped out before the list ever reaches you. A name you could not have used never makes it into the workbook.
Short pools, said plainly
If the pool runs out before your number does, the list arrives with what exists and the row tells you why — no padding, no quiet substitutions.
Pull more, get new people
Come back to a segment and pull again — the next batch is people you have not seen, not a re-run of the same names. Growing a list never means wading through people you already have.
More than an email
Every person carries their title, seniority, LinkedIn, city and state, then the company: domain, industry, size, founding year, revenue where published, and a description of what the business actually does. That last column is what makes a first line writable.
Business mailing addresses
Choose the direct-mail audience and the records carry a street address, city, state and postcode — business addresses, never residential — alongside a much higher rate of direct phone numbers. Old-school still works, and almost nothing else will build you the list.
Targeting by software
Reach companies by what they actually run — a storefront platform, an email tool, an analytics stack. One technology at a time, matched against the source's own list of over 1,600.
A workbook, not a dump
The client name, the audience and Claude's plain-English description of the criteria run across the top; every column filters; a Read Me sheet explains what each one means and how complete it is. It is the difference between handing a client a spreadsheet and handing them a piece of work.
Emails-only CSV
One column, deduplicated, ready to paste into a sending tool. Saves the step of stripping a workbook down before an import.
It keeps working if a source is down
Data providers go offline — several were during one week of testing. If the usual source cannot answer, the pull moves to another automatically and the list still arrives; the row tells you when that happened and what it changed.
- This builds business audiences. Reaching consumers is an advertising job, done with ads and your client's own customer data — not a purchased list, and this will not pretend otherwise.
- Emails are filtered to deliverable, but no source anywhere carries a “last verified” date — treat the first send as the confirmation, as you would with any list.
- Company descriptions come from the source, which truncates them at around 500 characters — some end mid-sentence. That is the data as published, not a trim on our side.
- Direct phone numbers are present on roughly a third of email-audience records, and far more of direct-mail ones. Mailing addresses are business addresses, never residential.
- Direct-mail lists are a one-time pull per set of criteria — the same criteria would return the same people. Change the geography or size band to reach new ones.
- Technology targeting matches one platform at a time. Asking for two widens the list rather than narrowing it, so it is offered as one.
- Small local trades — a two-truck contractor, a single-chair salon — are thinly covered by business databases everywhere. When a count comes back far below what you would expect, that is what you are seeing, and it will be said on the row.
- US-first by design. Contacts are filtered on where the person is, not where their company is headquartered.
See the count and a real sample before you decide anything.